SIOV metal oxide varistors - ホーム - EPCOS AG Principle of protection and characteristic...

SIOV metal oxide varistors

Selection procedure

Date: January 2018

© EPCOS AG 2018. Reproduction, publication and dissemination of this publication, enclosures hereto and theinformation contained therein without EPCOS' prior express consent is prohibited.

EPCOS AG is a TDK Group Company.

1 Selection procedure

1.1 Overvoltage types and sources

Overvoltages are distinguished according to where they originate.

1.1.1 Internal overvoltages

Internal overvoltages originate in the actual system which is to be protected, e.g. through

inductive load switching,

arcing,

direct coupling with higher voltage potential,

mutual inductive or capacitive interference between circuits,

electrostatic charge,

ESD.

With internal overvoltages the worst-case conditions can often be calculated or traced by a test

circuit. This enables the choice of overvoltage protective devices to be optimized.

1.1.2 External overvoltages

External overvoltages affect the system that is to be protected from the outside, e.g. as a result of

line interference,

strong electromagnetic fields,

lightning.

In most cases the waveform, amplitude and frequency of occurrence of these transients are not

known or, if so, only very vaguely. And this, of course, makes it difficult to design the appropriate

protective circuitry.

There have been attempts to define the overvoltage vulnerability of typical supply systems (e.g.

industrial, municipal, rural) so that the best possible protective device could be chosen for the pur-

pose. But the scale of local differences makes such an approach subject to uncertainty. So, for re-

liable protection against transients, a certain degree of “overdesign” must be considered.

Therefore the following figures for overvoltage in 230 V power lines can only be taken as rough

guidelines:

Amplitude up to 6 kV

Pulse duration 0.1 µs to 1 ms

Where varistors are operated directly on the line (i.e. without series resistor), normally the type

series S20 should be chosen. In systems with high exposure to transients (industrial, mountain lo-

cations) block varistors are to be preferred.

Requirements are stipulated in IEC 61000-4-X. Severity levels are specified in the respective

product standards.

Table 2 in chapter “Application notes” shows the selection of varistors for surge voltage loads ac-

cording to IEC 61000-4-5 as an example.

Selection procedure

Page 2 of 19Please read Important notesand Cautions and warnings.

1.2 Principle of protection and characteristic impedance

The principle of overvoltage protection by varistors is based on the series connection of voltage-

independent and voltage-dependent resistance. Use is made of the fact that every real voltage

source and thus every transient has a voltage-independent source impedance greater than zero.

This voltage-independent impedance Zsource in figure 1 can be the ohmic resistance of a cable or

the inductive reactance of a coil or the complex characteristic impedance of a transmission line.

If a transient occurs, current flows across Zsource and the varistor that, because vsource = Zsource · i,

causes a proportional voltage drop across the voltage-independent impedance. In contrast, the

voltage drop across the SIOV is almost independent of the current that flows.

Because

(equ. 8)

the voltage division ratio is shifted so that the overvoltage drops almost entirely across Zsource. The

circuit parallel to the varistor (voltage VSIOV) is protected.

Figure 1 Equivalent circuit in which Zsource symbolizes

the voltage-independent source impedance

Selection procedure

Page 3 of 19Please read Important notesand Cautions and warnings.

Figures 2 and 3 show the principle of overvoltage protection by varistors:

The intersection of the “load line” of the overvoltage with the V/I characteristic curve of the varistor

is the “operating point” of the overvoltage protection, i.e. surge current amplitude and protection

level.

Figure 2 Principle of overvoltage protection by varistors

Figure 3 Principle of overvoltage protection by varistors

The overvoltage ➀ is clamped to ➁ by a varistor.

Vop

Vsurge

Vclamp

Operating voltage

Superimposed surge voltage

Clamping voltage

For selection of the most suitable protective element, you must know the surge current waveform

that goes with the transient. This is often, and mistakenly, calculated by way of the (very small)

source impedance of the line at line frequency. This leads to current amplitudes of unrealistic pro-

portions. Here you must remember that typical surge current waves contain a large portion of fre-

quencies in the kHz and MHz range, at which the relatively high characteristic impedance of ca-

bles, leads, etc. determines the voltage/current ratio.

Selection procedure

Page 4 of 19Please read Important notesand Cautions and warnings.

Figure 4 shows approximate figures for the characteristic impedance of a supply line when there

are high-frequency overvoltages. For calculation purposes the characteristic impedance is nor-

mally taken as being 50 Ω. Artificial networks and surge generators are designed accordingly.

Figure 4 Impedance of a supply line for high-frequency overvoltages

Selection procedure

Page 5 of 19Please read Important notesand Cautions and warnings.

1.3 Areas of application for varistors

A wide selection of types is available to cover very different requirements for protective level and

load capability. Straightforward conditions of use and an attractive price/performance ratio have

made SIOVs from EPCOS successful in just about every area of electrical engineering and elec-

tronics. The table below summarizes them:

Telecommunications

Private branch exchanges

Telephone subscriber sets

Telephone pushbutton modules

Teleprinters

Answering sets

Power supply units

Transmitting systems

Fax machines

Modems

Cellular (mobile) phones

Cordless phones

Chargers

Car kits

Industrial controls

Telemetering systems

Remote control systems

Machine controls

Elevator controls

Alarm systems

Proximity switches

Lighting controls

Power supply units

Ground fault interrupters

Gas heating electronics

Electronic ballasts

LCDs

Power electronics

Bridge rectifiers

Brake rectifiers

Electric welding

Electric vehicles

Switch-mode power supplies

High-power current converters

DC/AC converters

Power semiconductors

Power engineering

Transformers

Inductors

Motor and generator windings

Electrical power meter

Automotive electronics

Central protection of

automotive electrical systems

Load-dump protection

Anti-skid brake systems

Trip recorders

Radios

Engine control units

Generator rectifiers

Central locking systems

Trip computers

Wiper motors

Power window systems

Airbag electronics

Carphones

Seat memories

Traffic lighting

Traffic signals

Runway lighting

Beacon lights

Medical engineering

Diagnostic equipment

Therapeutic equipment

Power supply units

Data systems

Data lines

Power supply units

Personal computers

Interfaces

ASIC resets

Microcontrollers

I/O ports

Keyboards

Handheld PCs

Stepped protection

Microelectronics

EMI/RFI suppression

EMP/NEMP protection

Entertainment electronics

Video sets

Television sets

Slide projectors

Power supply units

HIFI equipment

Set-top boxes

Household electronics

Washer controls

Dimmers

Lamps

Quartz clocks

Electric motor tools

Thermostats

Replacement of

Suppressor diodes

Diodes

Selection procedure

Page 6 of 19Please read Important notesand Cautions and warnings.

If semiconductor devices like diodes, thyristors and triacs are paralleled with SIOVs for protection,

they may do with lower reverse-voltage strength. This leads to a marked cost reduction and can

be the factor that really makes a circuit competitive.

1.4 Series and parallel connection

1.4.1 Series connection

SIOV varistors can be connected in series for more precise matching to uncommon voltage rat-

ings or for voltage ratings higher than those available. For this purpose the types selected should

be of the same series (i.e. same diameter). The maximum permissible operating voltage in series

configuration is produced by adding the maximum DC or AC voltages of the varistors.

1.4.2 Parallel connection

Metal oxide varistors can be connected in parallel to achieve higher current load capabilities or

higher energy absorption than can be obtained with single components. To this end, the intended

operating point in the surge current region (see chapter “General technical information”, section

1.5) must be taken into account.

1.4.2.1 Medium current region

Since the surge current is well below its maximum permissible value in this region, parallel con-

nection may only be used to increase energy absorption. The varistor has to absorb the energy of

currents that have a relatively low amplitude, but a high energy content due to their duration.

Example surge current i* = 1 A in figure 5:

In the worst case, 2 varistors may have been chosen for parallel connection with the first having a

V/I characteristic curve corresponding to the upper limits and the second having a V/I characteris-

tic curve corresponding to the lower limits of the tolerance band. From the region boundary a) one

can see that then a current of 1 mA flows through the first varistor and a current of 1 A flows

through the second varistor. The energy absorptions of the two varistors are in the same ratio.

This means that if unselected varistors are used in this current region, current distributions of up

to 1000:1 may render the parallel connection useless. To achieve the desired results, it is neces-

sary to match voltage and current to the intended operating point.

1.4.2.2 High-current region

In this region, the ohmic resistance of the zinc oxide causes a higher voltage drop across the

varistor that carries the higher surge current. Thus, the current distribution is shifted to the varistor

with the lower current. Region b) in figure 5 shows that in the worst case the current ratio is ap-

prox. 15 kA:40 kA, which is a considerably better result than in the medium operating region. Ac-

cordingly, parallel connection can increase the maximum permissible surge current for two block

varistors, e.g. from 40 kA to 55 kA for B40K275 varistors.

The graphical method in accordance with figure 5 can only provide guideline values, since the de-

viation of the individual varistors from the standard nonlinear values is not taken into considera-

tion. In practice, the individual varistors must be measured for the current region for which parallel

operation is envisaged. If this region is within the two upper decades of the maximum surge cur-

rent, the varistors should be measured at 1% of the maximum current to prevent the measure-

Selection procedure

Page 7 of 19Please read Important notesand Cautions and warnings.

ment itself reducing the service life of the varistor. Example: using B40K275, maximum permissi-

ble surge current 40 kA. The measurement should take place using 400 A with surge current

pulse 8/20 µs.

The effort required for measurements of this kind will make parallel connection an exception. The

possibility of using a single varistor with a higher load capacity should always be preferred, in this

example it would be a type from the LS50, B60 or B80 series.

Figure 5 Tolerance band of the SIOV-B40K275

1.5 Selection guide

The choice of a varistor involves three main steps:

Select varistors that are suitable for the operating voltage.

Determine the varistor that is most suitable for the intended application in terms of

a) surge current,

b) energy absorption,

c) average power dissipation,

(for a and b also estimating the number of repetitions).

Determine the maximum possible voltage rise on the selected varistor in case of overvoltage

and compare this to the electric strength of the component or circuit that is to be protected.

Selection procedure

Page 8 of 19Please read Important notesand Cautions and warnings.

To ensure proper identification of circuit and varistor data, the following distinction is made:

Maximum possible loading of varistor that is determined by the electrical specifications of the

intended location.

Identification: *

Maximum permissible loading of varistor that is given by its surge current and absorption capa-

bility.

Identification: max

(e.g. x*, xmax)

So the following must always apply:

i* ≤ imax (equ. 9)

W* ≤ Wmax (equ. 10)

P* ≤ Pmax (equ. 11)

1.5.1 Operating voltage

Maximum permissible AC and DC operating voltages are stated in the product tables for all varis-

tors. To obtain as low a protection level as possible, varistors must be selected whose maximum

permissible operating voltage equals or minimally exceeds the operating voltage of the applica-

tion.

Nonsinusoidal AC voltages are compared with the maximum permissible DC operating voltages

so that the peak or amplitude of the applied voltage does not exceed the maximum permissible

DC voltage.

Note:

Of course, you may also select any varistor with a higher permissible operating voltage. This pro-

cedure is used, for example, when it is more important to have an extremely low leakage current

than the lowest possible protection level. In addition, the service life of the varistor is increased.

Also the type for the highest operating voltage may be selected to reduce the number of types be-

ing used for different voltages.

1.5.2 Surge current

Definition of the maximum possible operating voltage in the previous step will have narrowed

down the choice of an optimum SIOV to the models of a voltage class (e.g. those whose designa-

tion ends in 275 for 230 V + 10% = 253 V). Then you check, with reference to the conditions of

the application, what kind of load the SIOV can be subjected to.

Determining the load on the varistor when limiting overvoltage means that you have to know the

surge current that is to be handled.

1.5.2.1 Predefined surge current

Often the surge current is predefined in specifications. After transformation into an equivalent rec-

tangular wave (figure 9) the suitable varistor type can be selected by the derating curves.

Selection procedure

Page 9 of 19Please read Important notesand Cautions and warnings.

1.5.2.2 Predefined voltage or network

If the voltage or a network is predefined, the surge current can be determined in one of the follow-

ing ways:

Simulation

Using the PSpice simulation models of the SIOV varistors, the surge current, waveform and ener-

gy content can be calculated without difficulty. In these models, the maximum surge current is de-

duced for the lower limit of the tolerance band, i.e. setting TOL = –10.

Test circuit

The amplitude and waveform of the surge current can be determined with the aid of a test circuit.

The dynamic processes for overvoltages require adapted measuring procedures.

Graphical method

As shown in figures 6 and 7, the overvoltage can be drawn into the V/I characteristic curve fields

as a load line (open circuit voltage, short circuit current). At the intersection of this “load line” with

the varistor curve selected to suit the operating voltage, the maximum protection level and the

corresponding surge current can be read off. The waveform and thus the energy content cannot

be determined by this method.

Since the V/I characteristic curves are drawn in a log-log representation, the “load line” in figure 7

is distorted to a curve.

Figure 6 Load line on linear scale

Selection procedure

Page 10 of 19Please read Important notesand Cautions and warnings.

Figure 7 V/I characteristic curves SIOV-S20 with the load line drawn in

for a surge current amplitude 4 kV with Zsource = 2 Ω

Selection procedure

Page 11 of 19Please read Important notesand Cautions and warnings.

Mathematic approximation

The surge current is determined solely from the source impedance of the surge voltage (Vs). By

subtracting the voltage drop across the varistor (from the V/I curve) you can approximate the

maximum surge current as follows:

See 4.2 for an example. (equ. 12)

Switching off inductive loads

If the transient problems are caused by switching off an inductor, the “surge current” can be esti-

mated as follows:

The current through an inductance cannot change abruptly, so, when switching off, a current of

the order of the operating current must flow across the varistor as an initial value and then decay

following an e function. The path taken by the current during this time is referred to as a flywheel

circuit (refer to chapters “Calculation examples”, “Switching off inductive loads”).

The time constant τ = L/R that can be calculated from the inductance and the resistance of the fly-

wheel circuit (including varistor resistance) shows how long the current requires to return to the

1/e part (approx. 37%) of its original value. According to theory, t is also the time that the flywheel

current must continue to flow at constant magnitude to transport the same charge as the decaying

current.

So the amplitude of the “surge current” is known, and its duration is approximately τ (figure 8).

τ depends on the value of the inductance and the resistances of the flywheel circuit, generally

therefore on the resistance of the coil and the varistor. The latter is, by definition, dependent on

voltage and thus also current and so, for a given current, it has to be calculated from the voltage

drop across the varistor (V/I characteristic).

L

RCu

RSIOV

[H]

[Ω]

[Ω]

Inductance

Coil resistance

SIOV resistance at operating current

(equ. 13)

RSIOV increases as current decreases. So τ is not constant either during a decay process.

This dependence can be ignored in such a calculation however.

For comparison with the derating curves of the current you can say that τ = tr (refer to chapters

“Calculation examples”, “Switching off inductive loads”).

Figure 8 Time constant of flywheel circuit

Selection procedure

Page 12 of 19Please read Important notesand Cautions and warnings.

1.5.2.3 Comparison: determined surge current / derating curve

The maximum permissible surge current of the SIOV depends on the duration of current flow and

the required number of repetitions. Taking these two parameters, it can be read from the derating

curves. It is compared to the maximum possible surge currents in the intended electrical environ-

ment of the varistor.

From the derating curves one can obtain maximum figures for rectangular surge current waves.

For correct comparison with these maximum permissible values, the real surge current wave (any

shape) has to be converted into an equivalent rectangular wave. This is best done graphically by

the “rectangle method” illustrated in figure 9.

Keeping the maximum value, you can change the surge current wave into a rectangle of the same

area. t*r is then the duration of the equivalent rectangular wave and is identical to the “pulse width”

in the derating curves. (The period T* is needed to calculate the average power dissipation result-

ing from periodic application of energy.)

Figure 9 Rectangle method

If the pulse load is known, then tr can be calculated using the following equation:

(equ. 14)

The duration of surge current waves is frequently specified using the 50% value of the trailing

edge (ref. figure 16 in “General technical information”). The decay pattern of such waves can be

represented by an exponential function.

Selection procedure

Page 13 of 19Please read Important notesand Cautions and warnings.

According to figure 10 and the equation derived from this,

(equ. 15)

the “equivalent rectangular wave” for such processes is found to be t*r = 1.43 Tr

Figure 10 Equivalent rectangular wave of an e-function

1.5.3 Energy absorption

When a surge current flows across the varistor, there will be absorption of energy. The amount of

energy to be absorbed by the varistor can generally be calculated by equation 6.

Calculation method

Often the energy absorption can be read directly from a storage oscilloscope or can be calculated

from the voltage/current curve using numerical methods. An example for W* = 100 J is shown in

figures 12 to 14 in chapter “Application notes”).

Simulation

Determination of the energy absorption by simulation (PSpice) is even more convenient.

Graphic method

Otherwise equation 6 can be solved graphically with sufficient accuracy by using the rectangle

method. i* (t) is converted as in figure 9 and multiplied by the highest voltage appearing on the

varistor according to equation 16:

[J]

[V]

(equ. 16)[A]

[s]

can either be derived from the V/I characteristic as the value matching , or likewise be deter-

mined with the aid of an oscilloscope as the maximum voltage drop across the varistor.

Selection procedure

Page 14 of 19Please read Important notesand Cautions and warnings.

Switching off inductive loads

If transients are caused by interrupting the current supply of an inductor, the worst-case principle

can be applied to calculate the necessary energy absorption of a varistor. The energy to be ab-

sorbed by the varistor cannot be greater than that stored in the inductor:

W* = ½ L i*2 [J] L

i*

[H]

[A](equ. 17)

This calculation will always include a safety margin because of losses in other components. Refer

to chapter “Calculation examples”, section 1.1.

Note:

When used for overvoltages caused by switching off inductive loads, varistors should always be

applied in freewheeling circuits as shown in “Calculation examples”, section 1.1 and figure 5 in

chapter “Application notes".

Discharging of capacitors

The statements made for inductors also apply for capacitances. This means that the load placed

on the varistors in many of the tests according to IEC 61000-4-X can be estimated.

Comparison: determined energy input / maximum permissible energy absorption

To check the selection requirement W* ≤ Wmax (equation 10), you have to determine the maximum

permissible energy absorption for the intended varistor. This can be calculated by equation 18 as

a function of the time the energy is applied (tr) and the number of repetitions from the derating

curves:

Wmax = vmax imax tr max (equ. 18)

vmax is derived from the V/I characteristic of the intended varistor type for the surge current imax.

tr max can be taken as being the same as t*r, because Wmax is to be calculated for the given time of

current flow.

1.5.4 Average power dissipation

The actual power dissipation of a varistor is composed of the basic dissipation P0 caused by the

operating voltage and, possibly, the average of periodic energy absorption. If metal oxide varis-

tors are chosen from the product tables in agreement with the maximum permissible operating

voltages, P0 will be negligible.

Periodic energy absorption produces an average power dissipation of:

W*

T*

v*

[J]

[s]

[V]

i*

t*r

P*

[A]

[s]

[W]

(equ. 19)

W* takes the value of a single absorption of energy.

T* is the period of figure 9.

Selection procedure

Page 15 of 19Please read Important notesand Cautions and warnings.

By solving this equation for T* it is possible to calculate the minimum time that must elapse before

energy is applied again without exceeding the maximum permissible average power dissipation of

the varistor:

W*

Pmax

[J]

[W]

Tmin [s](equ. 20)

Note:

Metal oxide varistors are not to be “operated” at Pmax. They are not suitable for “static” power dis-

sipation, e.g. voltage stabilization. There are other kinds of components, like zener diodes, de-

signed primarily for this kind of application, but with much lower surge current handling capability.

1.5.5 Maximum protection level

The maximum possible voltage rise in the event of a surge current is checked with the aid of the

V/I curves or PSpice models. This figure can be read directly from the curve for a given surge cur-

rent (for worst-case varistor tolerances). If the voltage value thus obtained is higher than accept-

able, the following possibilities may assist in reducing the protection level:

Choose a type with a larger disk diameter

The protection level is lower for the same surge current because the current density is reduced.

Better matching to the operating voltage by series connection

Example: 340 V AC

Here, according to the first step in selection, a standard SIOV with the end number “385” would

normally be chosen. But if two SIOVs with the end number “175” are connected in series, the

response of a 350 V varistor is obtained.

Choose a tighter tolerance band

A special type is introduced that only utilizes the bottom half of the standard tolerance band for

example. This would mean a drop in the protection level by approx. 10%.

Insert a series resistor

This reduces the amplitude of the surge current and thus the protection level of the varistor.

Note:

If the protection level obtained from the V/I curve is lower than required, you can change to a

varistor with a higher protection level, i.e. higher end number in its type designation. This has a fa-

vorable effect on load handling capability and operating life. The leakage current is further re-

duced. If necessary, the number of different types used can be reduced.

1.5.6 Selection by test circuit

The maximum permissible ratings of varistors refer to the amount of energy that will cause the

varistor voltage to change by maximum ±10%.

Figures 11 and 12 show typical curves for the change in varistor voltage of metal oxide varistors

when energy is repeatedly applied through a bipolar or unipolar load. You often find an increase

of a few percent to begin with, and for a unipolar load there are also polarization effects. This is

seen in figure 12 for the leakage current. Such phenomena have to be considered when interpret-

ing measured results.

Selection procedure

Page 16 of 19Please read Important notesand Cautions and warnings.

So, in test circuits, you start by determining the varistor voltage for every single type as accurately

as possible (at a defined temperature). It is advisable to check the change in varistor voltage from

time to time, making sure the temperature is the same. By extrapolation of the measured results

to the intersection with the –10% line, a guide value for the lifetime of varistors is obtained.

Figure 11 Typical curves for change in varistor voltage when metal oxide varistors are

repeatedly loaded

Figure 12 Typical polarization effect for unipolar loading of metal oxide varistors

Selection procedure

Page 17 of 19Please read Important notesand Cautions and warnings.

Figure 13 Typical polarization effects of leakage current for unipolar loading

of metal oxide varistors

Selection procedure

Page 18 of 19Please read Important notesand Cautions and warnings.

1.6 Questionnaire for selecting SIOVs

Table 1

Date . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Company . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Project . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Remarks

1. Max. operating voltage ... VAC or … VDC

Operating voltage including

all +tolerances

2.1 Surge current amplitude ............ kA

Waveform .............. s e.g. 8/20 µs, 10/1000 µs, 2 ms

No. of repetitions through lifetime ........ times

Minimum time between consecutive loads .............. s

and / or

2.2 Energy per load .............. J

Duration of energy input .............. s e.g. 8/20 µs, 10/1000 µs, 2 ms

No. of repetitions through lifetime ........ times

Minimum time between consecutive loads .............. s

and / or

2.3 Test standard .................. e.g.: IEC 61000-4-5

Test severity .................. e.g.: 3

Protection class .................. e.g.: 4

3. Required protection level

Reference current

... V

... A

4. Desired package Please mark

Disk type

Disk varistors in housing

ThermoFuse varistor (ETFV/ T)

EnergetiQ type

Block type

Strap type

Selection procedure

Page 19 of 19Please read Important notesand Cautions and warnings.